A Markov regime‐switching Cholesky GARCH model for directly estimating the dynamic of optimal hedge ratio
提出RSCHAR模型，通过Cholesky分解将原始序列转为正交因子并拟合机制转换动态，直接估计条件对冲比率，实证显示在方差降低、低阶矩和模型置信集上优于传统方法。
Abstract A Markov regime‐switching Cholesky GARCH (RSCHAR) model is proposed for directly estimating the optimal hedge ratio. The basic structure of RSCHAR is to transform the original series into a vector of orthogonal factors using Cholesky decomposition and fit these factors with regime‐switching dynamics. RSCHAR specifies directly the regime‐switching dynamic of conditional hedge ratio instead of recovering it indirectly from the estimated conditional covariance matrix. An estimation procedure is proposed for estimating RSCHAR. The empirical results reveal that RSCHAR exhibits superior effectiveness for multiple futures hedging based on the criterion of variance reduction, lower partial moment, and model confidence set.